Operating System Concepts

Chapter 3: Processes

• The text covers process concepts, process scheduling, operations on processes, interprocess communication (IPC), examples of IPC systems, and communication in client-server systems.

Important Terms

• Uniprogramming: RAM implies a single process.
• Multiprogramming: RAM implies multiple processes.
• Multitasking: RAM implies multiple processes that are belonging to different users.
• Multiprocessing: Implies multiple CPUs or microprocessors.
• Multithreading: Implies multiple threads inside a process.

Process Concept

• An operating system executes a variety of programs, including batch systems (jobs) and time-shared systems (user programs/tasks).
• The textbook uses the terms job and process almost interchangeably.
• A process is a program in execution, which must progress in a sequential fashion.
• A program is a passive entity stored on disk like an executable file, while a process is active.
• A program becomes a process when its executable file is loaded into the system memory.
• Execution of a program starts via GUI mouse clicks, command line entry of its name, and other methods.
• One program can have several processes, with multiple users executing the same program.

Process in Memory

• A process has multiple parts:
  • The text section, known as program code with the current activity, includes the program counter and processor registers.
  • The Stack contains temporary data like function parameters, return addresses, and local variables
  • The Data section contains global variables
  • The Heap contains memory dynamically allocated during run time.

Process Control Block (PCB)

• The PCB or task control block is associated to each process.
• Includes the process state (running, waiting, etc.) and the process identification number.
• Contains the program counter showing the location of instruction to next execute.
• It keeps CPU registers that contains the contents of all process-centric registers.
• It has CPU scheduling information like priorities, scheduling queue pointers.
• The PCB uses memory-management information which is the memory allocated to the process.
• Other information include accounting which has the CPU used, clock time elapsed since start, or time limits.
• The PCB stores the I/O status information for I/O devices allocated to process, list of open files

Process State transition

• As a process executes, it changes through different states.
• new: The process is being created.
• running: Instructions are being executed.
• waiting: The process is waiting for some event to occur.
• ready: The process is waiting to be assigned to a processor.
• terminated: The process has finished execution.
• Analogy for process is a hopsital - New admissions, waiting hall, Dr Checkup , laboratories etc..

Process Representation

• In Linux, a C structure "task struct" represents a process:
  • includes the PID (process identifier) and the state of the process.
  • contains scheduling information, a pointer to the parent process, and a list of children processes
  • Files shows a list of open files
  • mm struct shows the address space for this process

Threads

• Up to this poing processes had single threads of execution.
• Consider having multiple program counters per process.
• Multiple threads of control allow multiple locations to execute at once.
• Processes require storage for thread details and multiple program counters in the PCB.

Process Scheduling

• Process scheduling maximizes CPU use and quickly switches processes onto the CPU for time-sharing.
• The Process scheduler selects among available processes for next execution on CPU on a scheduling queue.
• Scheduling queues includes jobs that sets all the processes in the system.
• Ready queue is set of all processes residing in main memory, ready and waiting to execute.
  • A new process is initially put in the ready queue and waits there dispatch.
• Device queues are sets of processes waiting for an I/O device.
• Processes migrate among the various queues.

Schedulers

• Short-term scheduler (or CPU scheduler) selects which process should be executed next and allocates CPU.
  • Sometimes the only scheduler in a system, and it is invoked frequently in milliseconds, so it must be fast.
• Long-term scheduler (or job scheduler) selects which processes should be brought into the ready queue.
  • It is invoked infrequently in seconds or minutes, and it may be slow, controlling the degree of multiprogramming.
• Processes can be described as either I/O-bound that that spend more time doing I/O than computations, or
• CPU-bound processes spend more time doing computations with few very long CPU bursts.
• Long-term scheduler benefit good process mix.
• Time-sharing systems like UNIX and Microsoft Windows often do not have a long-term scheduler.
• A medium-term scheduler is added if the degree of multiple programming needs to decrease via swapping.

Context Switch

• When the CPU switches to another process, the system must save the state of the old process and load the saved state for the new process within a context switch.
• Overhead occurs with context-switch time, and the system does no useful work while switching.
• The more complex the OS and the PCB, the longer the context switch.
• Time is dependent on hardware support with some providing multiple sets of registers per CPU to load multiple contexts at once.

Operations on Processes

• An OS provides mechanisms for process creation and termination.

Process Creation

• A parent process creates child processes which in turn create other processes, forming a tree.
• Processes are identified and managed by a process identifier (pid).
• There are different resource sharing options that can occur.
• Including Parent and children sharing all resources.
• or Children share subset of parent's resources.
• Or Parent and child share no resources.
• Execution options has the parent and children executing concurrently.
• Or the parent should waits until children terminate

UNIX examples

• fork() system calls creates a new process.
• exec() system call is used after a fork() to replace the process' memory space with a new program.

Process Termination

• A process executes the last statement causing exit() system call.
• Then the process will return status data from the terminated child to parent via wait().
• The OS deallocates the processes resources.
• A parent may terminate the execution of children processes using the abort() system call for reasons like:
  • Child has exceeded allocated resources.
  • Task assigned to child is no longer required.
  • The parent is exiting, and the operating systems may not allow a child to continue if its parent terminates.
• Some operating systems do not release to exist child if its parent has been terminated.
• If a process terminates for cascading termination, then all its children (all the way down) must also be terminated.
• The parent process may wait for termination of a child process by using the wait() system call to return status information
• If no parent calling wait() on a terminated child process the program is called a zombie
• If a parent terminates and does not call wait then the process is an orphan.
• The init process is the root parent process (always has a pid of 1) for all user processes.

Multiprocess Architecture

• Chrome is an example of a multiprocess architecture.
• A design for running multiple processes with the 3 processes:
  • Browser process which will manages user interface, disk, and network I/O.
  • A Renderer process deals with rendering of web pages, HTML and Javascript.
    • new renderer is created for each website opened running in a sandbox to minimize the effect of security exploits.
  • Plug-in process for each type of plug-in

Interprocess Communication

• Processes within a system are independent or cooperating
• Independent processes can affect or be affected by the execution of another process.
• Cooperating processes is reverse.
• Reason to use:
  • Information sharing
  • Computation speedup
  • Modularity
  • Convenience
• Cooperating processes require interprocess communication (IPC) using two models: Shared and Message passing
• Shared Memory creates processes that exchange information by reading and writing data to the shared region. The communications under the control of the users processes.
• Message Passing takes place by means of messages exchange between the cooperating processes and easier to implement

Message Passing Implementation

• If processes P and Q communicate, establish communication link and exchange messages via send/receive.
  • Links can be through Direct or indirect communication, Synchronous or asynchronous, or Automatic or explicit buffering

Direct Communication:

• Proccessess must name or know each ther.
• use send(P, message) and recieve (Q,message)
• Links established automatically
• links associcated with one pair
• there is usually one link that is bi-directional

Indirect Communication

• Are directed and received from mailboxes
• Each has a unique ID
• Processes can communicate only sharing mailbox
• Link established with share mailbox
  • operations are send (A, message) and recieve (A,message)
  • is the size is either fixed or variable

###Synchronization during message passing

• Message passing may be either blocking (synchronous) or non-blocking asynchronous Blocking:
  • sennder is blocked during blocking sedn till message arrives
  • receiver blocked during blocking receive till message is ready

Non Blocking: - sender sends and continue - valid messagw/ or null message -

Buffering

• Uses queues of messages attached to the link of 3 ways implemented :
  • Has has no queues - zero capacity so must wait for reciever
  • Finite que length - of n messages -sender if full link
  • infinite length - unbounded capacity where sender never waits

Producer-Consumer Problem

• This focuses on a producer process that consumes information that is consumed by a consumer
• If is unrestricted like unbounded-buffer there is no limit on size
• If is constricted like bounded-buffer, it assumes a fixed size

Client-Server Systems

• These methods are for inter-process communication where processes are on different computers.
• Shared memory and Message passing can't work for this type of client-server communication. so methods used includes:
  • Sockets, Remote Procedure Calls, Pipes, and Remote Method Invocation (Java)

Sockets

• Is is endpoint for communication that starts at the message packet.
• Is between hosts
• the socket 161.25.19.8:1625 refers to port 1625 on host 161.25.19.8
• ports below 1024 are well known
• special adress is 127.0.0.1

Remote Procedure Calls

• RPC absstracts process between processes
• uses ports for service
• stubs are cleint proxxies
• calls marhsalls
• microsofts version called MIDL

Data representations with a EDL

• big endian and little endian
• Messages are can only be delivered once
• OS provides a matchmarker service to contect client and server

Pipes

• Acts as conduit allowing processes to communicate
• Communication can be uni/birectional
• used for parent-child relation
• Ordinary pipes are used only for within process creating. Parent create pipe to communicate to child.
• windows calls these anonymous pipes
• Named pipes are accessed with no parent-child relationship.